|Table of Contents|

The Military Order Recognition Technology in Battlefield Environment(PDF)

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

Issue:
2002年04期
Page:
438-441
Research Field:
Publishing date:
2002-08-30

Info

Title:
The Military Order Recognition Technology in Battlefield Environment
Author(s):
LuJianwei DingQinghai ① ZhuXueping XiongGang ②
Department of Science Research,Zhengzhou Air Defence Force Academy,Zhengzhou 450052)
Keywords:
military order parallel model combinat ion combinat ions method speech recognit ion
PACS:
E96
DOI:
-
Abstract:
Because of the part icularity in bat t lefield environment , environmental noise is a main obstruction in the military order recog nit ion technology. This paper addresses a method of automat ic speech recog nit ion in the presence of interfering noise ) parallel model combination ( PMC) . The theory of PMC method is discussed and used in noisy speech recognition in detail. And test it in three different kinds of noise and SNR using 48 ant iaircraf t fire command order. T he result show s that this method is of great pract ical use.

References:

1 Gales M J F, Young S. An improved approach to the hidden Markov model decomposition of speech and noise. Proceedings of Inter net Comfer ence on Acoucstics Speech and Sig nal Procession. Washington: IEEE USA, 1992. 233~ 236
2 Var ga A P, Moore R K. Hidden Mar cov model decomposition of speech and noise. Proceedings of Internet Comfer ence on Acoucstics Speech and Signal Pr ocession. Washington: IEEE USA,19901845~ 848
3 Gales M J F , Young S. Cepstral parameter compensation for HMM recog nition in noise. Speech Communication, 1993, 12: 231~ 239
4 Gales M J F , Young S. HMM recognition in noise using parallel model combination. Pr oceedings of Eurospeech. Berth: Electrical Publishing House, 1993
5 Martin F, Shikano K, Minami Y. Recognition of no isy speech by composition of hidden mar kov models. Proceedings of Eur ospeech. Ber th: Electrical Publishing House, 1993. 1 031~ 1 034
6 杨行峻, 迟惠开. 语音信号数字处理. 北京: 电子工业出版社, 1995

Memo

Memo:
-
Last Update: 2002-08-30